Counterfactual assessment of protected area avoided deforestation in Cambodia: Trends in effectiveness, spillover effects and the influence of establishment date
Benjamin Black, Brandon P. Anthony
Abstract
Whilst protected area (PA) networks remain a cornerstone of global conservation efforts, persistent underfunding alongside the increasing severity of environmental degradation has furthered interest in quantifying the ecological benefits they generate. However, the counterfactual assessment of measures of effectiveness, such as avoided deforestation relative to unprotected areas, are confounded by the non-random siting of PAs within landscapes, which creates differential resource extraction pressures. Quasi-experimental techniques such as statistical matching can account for differential pressures by selecting suitable control areas for comparison by minimising variance across a suite of environmental and socio-economic covariates. This study presents a national scale assessment of PA effectiveness for Cambodia, involving 46 PAs, employing nearest-neighbour propensity score matching using 8 covariates, with repeated random sampling, to quantify avoided deforestation in 3 different outcome periods between 2010 and 2018. The results demonstrate that, despite historic criticism, Cambodia’s PAs generated significant positive avoided deforestation in all periods, with forested land in PAs being as much as 12.5 % less likely to be deforested than matched unprotected forest. However, effect size declined over subsequent time periods likely due to a combination of rising deforestation pressure from agri-business concessions, followed by a large-scale expansion of the PA estate and institutional rearrangements. The effectiveness of PAs in Cambodia varied with respect to PA age and a significant positive spillover effect of PAs (maximum of 5.9 % reduction in deforestation probability) was observed in 2 km buffer zones of increasing distance to PA boundaries. Our findings are consistent with studies of other PA networks that report heterogeneous effects for these phenomena.